Effects of human land use and temperature on community dynamics in European forests

Climate change and human land use are thought to play a dominant role in the dynamics of European central-latitude forests in the Holocene. A wide range of mathematical and statistical models have been used to study the effects of these variables on forest dynamics, including physiologically-based simulations and phenomenological community models. However, for statistical analysis of pollen count data, compositional data analysis is particularly well suited, because pollen counts give only relative information. We studied the effects of changes in human land use and temperature on European central-latitude forest dynamics at 7 sites over most of the last 10ka, using a stochastic model for compositional dynamics of pollen count data. Our approach has a natural ecological interpretation in terms of relative proportional population growth rates, and does not require information on pollen production, dispersal, or deposition. We showed that the relative proportional population growth rates of Fagus and Picea were positively affected by intensified human land use, and that those of Tilia and Ulmus were negatively affected. Also, the relative proportional population growth rate of Fagus was negatively affected by increases in temperature above about 18∘C. Overall, the effects of temperature on the rate of change of forest composition were more important than those of human land use. Although there were aspects of dynamics, such as short-term oscillations, that our model did not capture, our approach is broadly applicable and founded on ecological principles, and gave results consistent with current thinking. •Climate change and human land use are drivers of Holocene forest dynamics.•We applied compositional stochastic differential equations to pollen counts.•Human land use favoured Fagus and Picea over Tilia and Ulmus.•Temperature had larger effects than human land use on rate of change.